2012/05/04 by Radhouane Boughamoura, Boughamoura, Radhouane, Lobna Hlaoua +3
Computer Science · #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Information Retrieval (cs.IR) #cs.AI #cs.IR
paper · pdf · doi:10.48550/arxiv.1205.0917
8th NCM: 2012 International Conference on Networked Computing and Advanced Information Management
arxiv created 2012/05/04 · arxiv updated 2012/05/07
Deep Web databases contain more than 90% of pertinent information of the Web. Despite their importance, users don't profit of this treasury. Many deep web services are offering competitive services in term of prices, quality of service, and facilities. As the number of services is growing rapidly, users have difficulty to ask many web services in the same time. In this paper, we imagine a system where users have the possibility to formulate one query using one query interface and then the system translates query to the rest of query interfaces. However, interfaces are created by designers in order to be interpreted visually by users, machines can not interpret query from a given interface. We propose a new approach which emulates capacity of interpretation of users and extracts query from deep web query interfaces. Our approach has proved good performances on two standard datasets.